Hey Team,
The Chicago AREDN team (CAREDN) is ready to take on supernodes. I'm standing up KV3T-CHI-SUPERNODE, and am in need of tunnels to make that a part of the supernode system.
This is an exciting upgrade because our local network has gotten very slow with thousands of nodes from all over the world constantly pinging one another. Turning off tunnels to the other networks makes the problems go away. I appreciate the AREDN team solving our network's problems before I complained about them!
Thanks,
Casey
KV3T
The Chicago AREDN team (CAREDN) is ready to take on supernodes. I'm standing up KV3T-CHI-SUPERNODE, and am in need of tunnels to make that a part of the supernode system.
This is an exciting upgrade because our local network has gotten very slow with thousands of nodes from all over the world constantly pinging one another. Turning off tunnels to the other networks makes the problems go away. I appreciate the AREDN team solving our network's problems before I complained about them!
Thanks,
Casey
KV3T
Since you are toward the middle of the country, plan on tunnels both to the west coast and east coast. I have been seeing a bunch of the supernode connections just being haphazardly done and it is causing bad routing decisions. For example I am sitting down here in northern Florida with a supernode and the other day trying to go to KV4ATV up in Panama City, Florida the routing went up Ohio -> San Francisco -> Alaska -> Tennessee -> Georgia and then finally to KV4ATV. Similarly there are times when my VOIP calls go all the way to the west coast just to be routed back to the east coast to connect to N2MH's PBX.
Much of these inefficiencies are being caused by several of the mesh networks that have links all over the place really need to start getting supernodes installed and consolidate all the tunnels through the supernodes (like you are doing). The other big reason is the way that the supernodes started and got interconnected (this is from what I can tell--I don't have personal knowledge of the true network development). Originally most of the supernode connections went back to W6BI and KN6PLV--both in California. There needs to be some more work on interconnecting tunnels on the east coast so that there are not these routes going all the way back to the west coast.
Have you looked at how the supernodes are currently connected up at https://arednmaps.xojs.org? Just click on the supernode selection in the box to just see the supernodes. One thing I noticed the other day was that the Wireguard tunnels are not showing up this map. I have a connection up to NC8Q in Ohio (which would also be a natual connection for you) but it is not showing up. NC8Q desperartly needs to get a connection to the east coast, so maybe I will go email him with the suggestion.
I would also look at the supernode in Oklahoma or North Texas for a connection and then also decide on east coast and west coast connections. Then you will be pretty well set.
If you would like a connection down here to Florida let me know, but I think the better solution for you would be NC8Q and then he would route down this way for you.
73
Gerard, WTØF
Hi, Casey:
This is an offer for a supernode tunnel to Ohio.
Hi, Gerard:
NC8Q-OH-SUPERNODE 'desperately' needs to upgrade...
;-)
However, it is an x86 and I think I will need to manually re-enter the tunnel credentials after
loading a new OS. I will likely create a text file and do many iterations of 'copy-n-paste'.
Egad!
73, Chuck
73
Gerard, WT0F
root@NC8Q-OH-SUPERNODE:/# cat /var/vtun/vtund.conf
options {
port 5526;
timeout 60;
syslog daemon;
ppp /usr/sbin/pppd;
ifconfig /sbin/ifconfig;
route /sbin/route;
firewall /usr/sbin/nft;
ip /sbin/ip;
}
default {
persist yes;
}
root@NC8Q-OH-SUPERNODE:/# cat /var/vtun/vtundsrv.conf
options {
port 5526;
timeout 60;
syslog daemon;
ppp /usr/sbin/pppd;
ifconfig /sbin/ifconfig;
route /sbin/route;
firewall /usr/sbin/nft;
ip /sbin/ip;
}
root@NC8Q-OH-SUPERNODE:/# ls /var/wireguard/
root@NC8Q-OH-SUPERNODE:/#
Hi, Gerard:
No tunnel credentials there.
73, Chuck
Actually I was just poking around and you probably want to grab everything in /etc/config.mesh. You can then put it on the new node and execute /usr/local/bin/node-setup to reconfigure the node.
73
Gerard, WTØF
I just found /etc/config.mesh/wireguard and came here to tell you! :-|
Thanks for the tip on executing node-setup.
If this works I can upgrade and retain the wireguard keys.
Else, I'll be emailing my tunnel clients with new keys. :-(
73, Chuck
73
Gerard, WTØF
Its purpose to let the local Southwest Ohio NC8Q-SUPERNODE only have 1 supernode link to keep up with
The NC8Q-SUPERNODE is a hAP-ac2 and was running (IMHO) a HIGH 'load average'.
Letting it link with only 1 supernode keeps the 'load average' lower.
KN6PLV-BAM-SUPERNODE ; If some else has a more reliable link to California, I can disable this link.
KV3T-CHI-SUPERNODE ; Illinois. Established 2024-02-04
N3FUD-PA-SUPERNODE ; Pennsylvania
N5MXI-STX-SUPERNODE ; South Texas
NC8Q-SUPERNODE ; This is the local Miami Valley Microwave Alliance (70+ nodes) in Southwest Ohio
WT0F-NFL-SUPERNODE ; Northern Florida
I want this 'supernode' networking to work well and one way to tell is via VoIP calls.
If I can hear the latency during a conversation, I will test other routes to find the lowest latency.
What is invisible is the internet hops between our ISPs.
I can run 'mtr' to watch the end-to-end latency.
I think I/we can adjust the metrics to avoid route 'flapping'.
73, Chuck
That being said, you want to contact Casey (casey@caseydiers.com), KV3T, or (I think) Raphael (I don't have email at my fingertips), AC9WM, about arranging a tunnel. Each of them have a supernode up in the Chicago area and should be able to get you connected.
73
Gerard, WT0F